The perception of linear self-motion in response to combinations of visual and physical motion cues

During self-motion our senses transduce several partially redundant cues into a single perception of self-motion having a magnitude and direction. How are the senses combined to achieve this? This thesis examines this complex question using passive linear motion with different motion profiles while varying visual and physical motion cues.; Self motion was simulated with various cues presented either uni-modal or multi-modal. By combining stimuli corresponding to different motions at the same time, their various contributions could be teased apart. Independent control of the components of the visual stimulus was achieved by using high resolution virtual reality simulations in a helmet mounted display or projected onto large screens. Subjects were provided with physical motion cues by having them ride on a physically moving cart. The perceived magnitude of motion was assessed by presenting subjects with a visual distance and then having them compare their perception of motion to this distance. Perceived heading direction was measured by adjusting an unseen pointer.; Rather than matching their self-motion accurately to the target distance, subjects made systematic errors. Their perceived magnitude of self-motion varied with the many aspects of the stimulus. Low acceleration movements (below approximately 0.05m·s-2 including constant velocity) when defined by vision alone were consistently and systematically overestimated. Higher accelerations defined by vision alone tended to be judged more veridically. The response to visual stimulation also depended on other aspects of the visual display such as the presence and combination of disparity cues and the size of the visual field. Perceived self-motion using physical motion cues was consistently overestimated for both low and high accelerations.; The perceived magnitude of displacement for combined visual and physical motions was dominated by the physical motion component, but only by the component of the physical motion in the visually defined direction.; Perceived heading direction was remarkably accurate in response to visual stimulation alone and was unaffected by concurrent physical motion even if the physical motion was in a different direction. Magnitude responses varied with simulated direction of optic flow even when subjects received visual and physical motion cues.; A linear additive model was used to model the combination of visual and physical motion in the straight-ahead direction. The magnitude of the perceived displacement caused by visual motion alone was modelled by the output of two filters tuned to radial and lamellar optic flow respectively. The weighted output of this model was combined with the weighted output of a model of the vestibular system to simulate the response to combinations of visual and physical motions. It was found not be necessary to include a nonlinear conflict detection stage in assigning these weights.